Visulaization
XiaZhao
4/17/2021
Requires
library(tidyverse)
library(ggthemes)
library(ggsci)
library(ggpubr)
library(survminer)
library(survival)
library(survivalROC)
library(reshape2)
library(data.table)
library(ggExtra)
library(cowplot)
library(ComplexHeatmap)
library(scico)
library(colorspace)
library(RColorBrewer)
library(lubridate)
library(tableone)
library(kableExtra)
source("../R_function/colors.R")
source("../R_function/surv_plot.R")
theme_set(theme_cowplot())
"%ni%" <- Negate("%in%")
options(stringsAsFactors = F)BMI and FBratio
df<-fread("../Data/Data/Phylum_cli_111samples.csv",data.table = F)
df$FBratio<-df$Firmicutes/df$Bacteroidetes
df$FBratio_g<-ifelse(df$FBratio>=median(df$FBratio),"High","Low")
data<-subset(df,Site=="Stool"&Response!="NE"&Cycle=="BL")
par(mfrow=c(1,2))
hist(data$BMI,main="Frequence of BMI",xlab = "BMI")
hist(data$FBratio,main="Frequence of FBratio",xlab = "FBratio")
data$BMI_g<-ifelse(data$BMI>25,"High","Low")
data$FBratio_g<-ifelse(data$FBratio>median(data$FBratio),"High","Low")
fit<-survfit(Surv(PFStime,PFS) ~ BMI_g,
data = data)
fit## Call: survfit(formula = Surv(PFStime, PFS) ~ BMI_g, data = data)
##
## n events median 0.95LCL 0.95UCL
## BMI_g=High 10 8 3.28 2.20 NA
## BMI_g=Low 22 21 1.97 1.87 4.2ggsurvplot(fit, data=data,xlab = "Time(months)",
censor.size=0.5, size = 0.5,
conf.int = T,tables.theme = theme_few(base_size = 6),
linetype = "strata",
legend.title = "",palette = c("black","red"),
risk.table = T,
#legend = c(0.84, 0.8),
pval = TRUE,pval.size = 3,
pval.coord=c(0.8,0.2),pval.method=F,
pval.method.coord=c(0.05,0.3),
ggtheme = theme_minimal() +
theme(line = element_line(size = 0.1),
text = element_text(size = 6)),
risk.table.col = "strata",
surv.median.line = "hv",
risk.table.y.text.col = T,
risk.table.y.text = FALSE )
p1<-ggscatter(subset(df,Cycle=="BL"&Response!="NE"&FBratio<10), x = "FBratio", y = "BMI",size=0.5,mean.point = T,
color = "Site", add.params = list(c(size=0.5,color="Site")),
add = "reg.line", conf.int = TRUE)+
stat_cor(label.x = 0.3,aes(color=Site))+
theme_few(base_size = 8)+
scale_color_aaas()+
scale_color_manual(values = col11)## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.p2<-ggscatter(subset(df,Cycle=="BL"&Response!="NE"), x = "Firmicutes", y = "Bacteroidetes",size=0.5,mean.point = T,
color = "Site", add.params = list(c(size=0.5,color="Site")),
add = "reg.line", conf.int = TRUE)+
stat_cor(label.x = 0.2,aes(color=Site))+
theme_few(base_size = 8)+
scale_color_aaas()
p3<-ggstatsplot::ggbarstats(data = data,x=Response,ggtheme = ggplot2::theme_bw(base_size=8),bias.correct = T,
y =FBratio_g,subtitle = F,results.subtitle=F,
ggstatsplot.layer = FALSE,
legend.position="right",
messages = FALSE,
package = "ggsci",
palette = "default_nejm",
main = Response, nboot = 100,
legend.title = "Response")## Registered S3 methods overwritten by 'lme4':
## method from
## cooks.distance.influence.merMod car
## influence.merMod car
## dfbeta.influence.merMod car
## dfbetas.influence.merMod carplot_grid(p1,p2, p3,labels = c("A","B","C"), ncol =3, nrow = 1)## `geom_smooth()` using formula 'y ~ x'## `geom_smooth()` using formula 'y ~ x'
## Treated-related aderse events
df <- fread("../Data/Data/paired_BL_treat_16patients.csv", data.table = F)
df$Hand_food_syndrom <- as.factor(df$Hand_food_syndrom)
df$Hand_food_syndrom_g <- ifelse(df$Hand_food_syndrom %in% c("0", "1"), "no", "yes")
df$Rash <- as.factor(df$Rash)
df$Rash_g <- ifelse(df$Rash == "0", "no", "yes")
df$Fever <- as.factor(df$Fever)
df$Fever_g <- ifelse(df$Fever == "0", "no", "yes")
df$Diarrhea <- as.factor(df$Diarrhea)
df$Diarrhea_g <- ifelse(df$Diarrhea == "0", "no", "yes")
df_treat <- subset(df, Group == "Treat")
# List of ggsurvplots
require("survminer")
splots <- list()
fit_PFS <- survfit(Surv(PFStime, PFS) ~ Hand_food_syndrom_g, data = df_treat)
fit_PFS
fit_OS <- survfit(Surv(OStime, OS) ~ Hand_food_syndrom_g, data = df_treat)
fit_OS
splots[[1]] <- surv_plot(fit_PFS, df_treat, colors = c("darkgreen", "darkorange"),
title = "HandFoodSyndrom_PFS")
## Loading required package: prodlim
splots[[2]] <- surv_plot(fit_OS, df_treat, colors = c("black", "red"), title = "HandFoodSyndrom_OS")
fit_PFS <- survfit(Surv(PFStime, PFS) ~ Rash_g, data = df_treat)
fit_PFS
fit_OS <- survfit(Surv(OStime, OS) ~ Rash_g, data = df_treat)
fit_OS
splots[[3]] <- surv_plot(fit_PFS, df_treat, colors = c("darkgreen", "darkorange"),
title = "Rash_PFS")
splots[[4]] <- surv_plot(fit_OS, df_treat, colors = c("black", "red"), title = "Rash_OS")
fit_PFS <- survfit(Surv(PFStime, PFS) ~ Fever_g, data = df_treat)
fit_PFS
fit_OS <- survfit(Surv(OStime, OS) ~ Fever_g, data = df_treat)
fit_OS
splots[[5]] <- surv_plot(fit_PFS, df_treat, colors = c("darkgreen", "darkorange"),
title = "Fever_PFS")
splots[[6]] <- surv_plot(fit_OS, df_treat, colors = c("black", "red"), title = "Fever_OS")
fit_PFS <- survfit(Surv(PFStime, PFS) ~ Diarrhea_g, data = df_treat)
fit_PFS
fit_OS <- survfit(Surv(OStime, OS) ~ Diarrhea_g, data = df_treat)
fit_OS
splots[[7]] <- surv_plot(fit_PFS, df_treat, colors = c("darkgreen", "darkorange"),
title = "Diarrhea_PFS")
splots[[8]] <- surv_plot(fit_OS, df_treat, colors = c("black", "red"), title = "Diarrhea_OS")
arrange_ggsurvplots(splots, print = TRUE, ncol = 4, nrow = 2)
bar1 <- ggplot(df, aes(Group, Desulfovibrionaceae, fill = Response)) + geom_boxplot() +
geom_line(aes(group = patientID, color = Response, size = Desulfovibrionaceae),
alpha = 0.5) + geom_point(aes(size = Desulfovibrionaceae), color = "darkblue",
alpha = 0.5) + theme_few(base_size = 8) + stat_compare_means(label = "p.signif") +
scale_fill_d3() + theme(legend.key = element_blank(), axis.title.x = element_blank())
bar2 <- ggplot(df, aes(Group, Desulfovibrionaceae, fill = Diarrhea_g)) + geom_boxplot() +
geom_line(aes(group = patientID, color = Diarrhea_g, size = Desulfovibrionaceae),
alpha = 0.5) + geom_point(aes(size = Desulfovibrionaceae), color = "darkblue",
alpha = 0.5) + theme_few(base_size = 8) + stat_compare_means(label = "p.signif") +
scale_fill_jama() + theme(legend.key = element_blank(), axis.title.x = element_blank())
## Call: survfit(formula = Surv(PFStime, PFS) ~ Hand_food_syndrom_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Hand_food_syndrom_g=no 7 7 2.3 1.97 NA
## Hand_food_syndrom_g=yes 9 7 4.2 2.20 NA
## Call: survfit(formula = Surv(OStime, OS) ~ Hand_food_syndrom_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Hand_food_syndrom_g=no 7 2 NA 5.17 NA
## Hand_food_syndrom_g=yes 9 3 15.5 10.33 NA
## Call: survfit(formula = Surv(PFStime, PFS) ~ Rash_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Rash_g=no 11 10 3.80 2.03 NA
## Rash_g=yes 5 4 4.23 2.20 NA
## Call: survfit(formula = Surv(OStime, OS) ~ Rash_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Rash_g=no 11 3 10.3 10.3 NA
## Rash_g=yes 5 2 15.5 15.5 NA
## Call: survfit(formula = Surv(PFStime, PFS) ~ Fever_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Fever_g=no 14 12 3.05 2.03 NA
## Fever_g=yes 2 2 5.28 4.23 NA
## Call: survfit(formula = Surv(OStime, OS) ~ Fever_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Fever_g=no 14 4 15.5 15.5 NA
## Fever_g=yes 2 1 10.3 NA NA
## Call: survfit(formula = Surv(PFStime, PFS) ~ Diarrhea_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Diarrhea_g=no 13 11 4.23 2.30 NA
## Diarrhea_g=yes 3 3 1.97 1.87 NA
## Call: survfit(formula = Surv(OStime, OS) ~ Diarrhea_g, data = df_treat)
##
## n events median 0.95LCL 0.95UCL
## Diarrhea_g=no 13 4 15.5 10.3 NA
## Diarrhea_g=yes 3 1 NA 3.9 NAplot_grid(bar1, bar2, labels = c("A", "B"), ncol = 2, nrow = 1)